New options in Gemini 2.5
Native audio output and reside API enhancements
as we speak, Live API is introducing preview variations of audiovisual enter and native voice output dialogs, so you possibly can construct conversational experiences instantly with Gemini, that are extra pure and expressive.
Customers also can regulate their tone, accent, and talking type. For instance, you possibly can inform your mannequin to make use of a dramatic voice when telling a narrative. We additionally assist the usage of instruments and permit us to go looking in your behalf.
You may check out an preliminary set of options, together with:
- emotional dialogue. The mannequin detects feelings in your voice and responds appropriately.
- With proactive audio, the mannequin ignores background conversations and is aware of when to reply.
- Pondering with Reside API. The mannequin leverages Gemini’s considering capabilities to assist extra complicated duties.
We’re additionally releasing a brand new preview of text-to-speech for two.5 Professional and a pair of.5 Flash. For the primary time, they assist a number of audio system and allow two-voice text-to-speech through native audio output.
Like native audio dialogue, text-to-speech is expressive and may seize probably the most refined nuances, similar to whispers. Works with over 24 languages and seamlessly switches between them.
